Topical administration of DPP-IV inhibitors prevents retinal neurodegeneration in experimental diabetes
The main aims of the present study were: (1) to assess the expression and content of dipeptidyl peptidase IV (DPP-IV) in human and db/db mouse retinas, and in human vitreous fluid; and (2) to determine whether the topical administration of the DPP-IV inhibitors (DPP-IVi) would prevent retinal neurodegeneration and vascular leakage in db/db mice by reducing endogenous glucagon-like peptide 1 (GLP-1) degradation.
To assess the expression and content of DPP-IV, human samples of vitreous fluid and retinas were obtained from participants with type 2 diabetes (n = 8) and age-matched non-diabetic individuals (n = 8), as well as from db/db (n = 72) and db/+ (n = 28) mice. The interventional study, which included 72 db/db mice, consisted of the topical administration (eye drops) of saxagliptin, sitagliptin or vehicle for 14 days. DPP-IV mRNA levels were assessed by RT-PCR, and protein content was measured by ELISA or western blotting. GLP-1 was assessed by immunofluorescence, and its downstream effector exchange protein activated by cAMP-1 (EPAC-1) was used as a measure of GLP-1 receptor activation. Retinal analyses were performed in vivo by electroretinography and ex vivo by RT-PCR (Epac-1, Iba-1 [also known as Aif1]), western blotting (EPAC-1, glial fibrillar acidic protein [GFAP], glutamate−aspartate transporter [GLAST]) and immunofluorescence measurements (GLP-1, GFAP, ionised calcium binding adaptor molecule 1 [IBA-1], TUNEL, GLAST, albumin and collagen IV). Glutamate was quantified by HPLC. In addition, vascular leakage was examined by the Evans Blue method.
DPP-IV was present in human vitreous fluid but in a range 100-fold less than in plasma. Both mRNA levels and protein content were much lower in the retina than in the liver or bowel, but were significantly higher in retinal pigment epithelium (RPE) from diabetic donors in comparison to non-diabetic donors (p < 0.05). Topical treatment with DPP-IVi prevented glial activation, apoptosis and vascular leakage induced by diabetes in db/db mice (p < 0.05). Moreover, it also significantly prevented diabetes-induced functional abnormalities in the electroretinogram. A significant increase of both GLP-1 and EPAC-1 was found after treatment with DPP-IVi (p < 0.05). Furthermore, GLAST downregulation induced by diabetes was prevented, resulting in a significant reduction of extracellular glutamate concentrations. All these effects were observed without any changes in blood glucose levels.
The topical administration of DPP-IVi is effective in preventing neurodegeneration and vascular leakage in the diabetic retina. These effects can be attributed to an enhancement of GLP-1, but other mechanisms unrelated to the prevention of GLP-1 degradation cannot be ruled out.
KeywordsDiabetic retinopathy DPP-IV inhibitors Experimental diabetes GLP-1 Retinal neurodegeneration
Dipeptidyl peptidase IV
Dipeptidyl peptidase IV inhibitor
Exchange protein activated by cAMP-1
Factor II receptor-like 1
Glial fibrillar acidic protein
Glucagon-like peptide 1
Glucagon-like peptide 1 receptor
Ionised calcium binding adaptor molecule 1
Interleukin-1 receptor antagonist
Retinal pigment epithelium
Cardiovascular and Other Long-term Outcomes with Semaglutide in Subjects with Type 2 Diabetes
Vascular endothelial growth factor
Although diabetic retinopathy has been classically considered to be a microcirculatory disease of the retina, there is emerging evidence to suggest that retinal neurodegeneration is an early event in the pathogenesis of diabetic retinopathy that could participate in the development of microvascular abnormalities [1, 2, 3, 4, 5, 6]. In fact, neural apoptosis and reactive gliosis, the hallmarks of retinal neurodegeneration, have already been observed in diabetic donors without microcirculatory abnormalities [7, 8]. There are several possible approaches to treating neurodegeneration (i.e. blocking the glutamate signalling pathway, replacing downregulated neuroprotective factors and improving the neurovascular coupling function), which have been reviewed .
We have recently found that the glucagon-like peptide 1 (GLP-1) receptor (GLP-1R) is abundantly produced in the human retina, and that the topical administration of GLP-1R agonists is able to prevent the neurodegenerative process (reactive gliosis and apoptosis) in the retina in a mouse model of spontaneous type 2 diabetes (db/db mouse) that also expresses GLP-1R in the retina . In addition, we found the presence of GLP-1 in the human retina, but both mRNA levels and protein content were significantly lower in retinas from diabetic in comparison with control individuals.
Furthermore, the effect of GLP-1 administered via eye drops was very similar to that of GLP-1R analogues. This was an unexpected finding because native GLP-1 is extremely susceptible to the catalytic activity of the enzyme dipeptidyl peptidase IV (DPP-IV), which cleaves off the two NH2-terminal amino acids. Consequently, GLP-1 rapidly degrades, showing a half-life in plasma of 1–2 min . However, the presence of DPP-IV in the human retina and vitreous humour has never been examined. In addition, the potential capacity of DPP-IV inhibitors (DPP-IVi) administered by the topical route (i.e. as eye drops) in preventing retinal neurodegeneration induced by diabetes remains to be investigated.
On this basis, the aims of the present study were: (1) to assess the expression and content of DPP-IV in the human retina and vitreous fluid; (2) to determine whether the topical administration (eye drops) of the DPP-IVi saxagliptin and sitagliptin would reduce GLP-1 degradation, thus preventing neurodegeneration and vascular leakage in db/db mice; and (3) to investigate whether, as occurs with GLP-1, the amelioration of glutamate-mediated excitotoxicity would be among the mechanisms involved in its potential beneficial effects.
The study included eight participants with type 2 diabetes who had undergone vitrectomy owing to proliferative diabetic retinopathy. Eight age-matched non-diabetic participants with other conditions requiring vitrectomy (epiretinal membrane or macular hole) served as a control group. Undiluted vitreous samples were obtained at the onset of vitrectomy. Vitreous samples from individuals who had received intravitreal injections of anti-vascular endothelial growth factor (VEGF) agents or corticosteroids 6 months prior to vitrectomy were rejected.
All individuals included in the study were recruited from Vall d’Hebron University Hospital, Barcelona, Spain. The research followed the tenets of the Declaration of Helsinki. The protocol for sample collection was approved by the hospital ethics committee, and the study participants gave informed consent.
Retinas were obtained from the tissue bank of our centre (Banc de Sang i Teixits Hospital Universitari Vall d’Hebron). Eight diabetic and eight non-diabetic donors matched by age and sex were included in the study. Fundoscopic examination of diabetic donors performed the year before death revealed the absence of microvascular abnormalities or mild non-proliferative diabetic retinopathy. Accordingly, these selected retinas had not previously been treated with intravitreous corticosteroids or anti-VEGF injections. The general characteristics of participants with diabetes and non-diabetic control individuals, the cause of death and glucose-lowering treatment are summarised in electronic supplementary material [ESM] Table 1.
One eye-cup was harvested in order to separate the neuroretina from the retinal pigment epithelium (RPE), and both tissues were immediately frozen with liquid nitrogen and stored at −80°C. Tissues derived from this eye-cup were used for the studies of gene expression and protein measurement. The other eye-cup was also harvested, and both the RPE and neuroretina were soaked in paraffin and used for immunohistochemical studies. The time period from death to eye enucleation was 3.7 ± 1.5 h.
The procedure for eye-cup donation and for the handling of this biological material was regulated by the protocol of donations of the tissue bank of our centre and was approved by the ethical committee.
RNA extraction and quantitative RT-PCR
Retinal sections from the eight diabetic and eight non-diabetic donors were deparaffinised in xylol and were rehydrated in graded ethanol. To eliminate autofluorescence, slides were washed in potassium permanganate. The methodology used is detailed in the ESM Methods.
DPP-IV was assessed in plasma, vitreous and retinal extracts (RPE and neuroretina) by a quantitative sandwich enzyme immunoassay (R&D Systems, Minneapolis, MN, USA) with a lower limit of detection of 0.016 ng/ml. The DPP-IV concentration in vitreous fluid and retinal extracts was normalised to total protein levels. Total protein concentration in the retinal and vitreous samples was obtained by bicinchoninic acid (BCA) assay (Thermo Scientific, Rockford, IL, USA).
Two separate studies were performed, one testing saxagliptin and the other testing sitagliptin. A total of 36 male db/db (BKS.Cg-Dock7m +/+Leprdb/J) mice aged 8 weeks were purchased (Charles River Laboratories, Calco, Italy) for the saxagliptin study. In addition, 14 non-diabetic (db/+; BKS.Cg-Dock7m +/+Leprdb/+) mice matched by age were used as a control group.
Animals were acclimated 2 weeks prior to the study under tight environmental conditions of humidity (60%), temperature (20°C) and cycles of 12 h/12 h light/darkness. They had free access to filtered water and ad libitum food (ENVIGO Global Diet Complete Feed for Rodents, Mucedola, Milan, Italy). Then, in order to minimize variability, animals were randomly housed (block randomisation) in groups of four mice per cage. Each cage contained nesting material and absorbent bedding (BioFresh Performance Bedding 1/8″ Pelleted Cellulose, Absorption Corp, Ferndale, WA, USA).
When the mice were aged 10 weeks, saxagliptin or vehicle eye drops were administered directly onto the superior corneal surface of each eye using a syringe. One drop (5 μl) of saxagliptin (1 mmol/l; 31.5 mg/ml), or vehicle (5 μl PBS, pH 7.4) was administered twice daily for 14 days. On day 15, one drop of either saxagliptin or vehicle was administered to the animals’ eyes 1 h prior to euthanasia.
The experimental design used for sitagliptin was the same as for saxagliptin, and a further 36 male db/db mice and 14 db/+ mice were used (supplier details the same). For the sitagliptin study, the concentration in the eye drops was 50 mg/ml sitagliptin.
On day 15, the mice were anaesthetised (1 ml ketamine/0.3 ml xylazine) and transcardially perfused. The animals were euthanised by cervical dislocation, and the eyes were immediately enucleated. This study was approved by the Animal Care and Use Committee of VHIR (Vall d’Hebron Research Institute). All the experiments were performed in accordance with the tenets of the European Community (86/609/CEE) and the Association for Research in Vision and Ophthalmology (ARVO).
Full-field electroretinogram recordings were measured using the Ganzfeld ERG platform (Phoenix Research Laboratories, Pleasanton, CA, USA), as reported elsewhere  and following the recommendations of the International Society for Clinical Electrophysiology of Vision .
The eyes were immediately enucleated, and the neuroretinas were separated (see ESM Methods). For immunohistochemical analysis, five sections of retina were obtained per animal. Images were acquired with a confocal laser scanning microscope (FV1000; Olympus, Hamburg, Germany). Five fields (three corresponding to the central and two to the peripheral retina) from each section were analysed. The same locations and number of fields were measured in all retinas.
RNA isolation and quantitative RT-PCR assay
Proteins were extracted from the neuroretinas in 80 μl of lysis buffer (RIPA buffer: phenylmethanesulfonylfluoride [PMSF], 1 mmol/l; Na3VO4, 2 mmol/l; NaF, 100 mmol/l) and 1X protease inhibitor cocktail (Sigma, St Louis, MO, USA) (see ESM Methods).
Immunohistochemical analysis for glial activation assessment
Glial activation was evaluated by fluorescence microscopy using specific antibodies against glial fibrillar acidic protein (GFAP) following the methodology described elsewhere . To evaluate the degree of glial activation, we used a scoring system based on the extent of GFAP staining  that has previously been used by our group [9, 11, 13].
Immunohistochemical analysis for apoptosis assessment
Apoptosis was evaluated using the TUNEL method, as previously described  (see ESM Methods). The results are presented as the percentage of TUNEL-positive cells with respect to the Hoechst-stained cells obtained by Image J software (National Institutes of Health, Bethesda, MD, USA).
Measurement of retinal vascular permeability
Retinal vascular permeability was examined by assessing albumin leakage from the blood vessels into the retina using double-immunostaining for albumin and collagen IV (in vitro) and the Evans Blue albumin method (ex vivo).
In the process of double-immunostaining for albumin and collagen IV, labelling with collagen IV localised the basement membrane of vessels and allowed us to evaluate the integrity of the blood–retinal barrier (BRB) by examining whether the albumin molecules remain confined within the retinal vessels or were in contrast extravasated (see ESM Methods).
The Evans Blue albumin method was used as previously described with some modifications [14, 15, 16, 17] (see ESM Methods). Digital images from different random fields of all retinas were acquired using a confocal laser scanning microscope (FV1000; Olympus) at ×20 using the 561 nm laser line, and each image was recorded with identical beam intensity at a size of 1024 pixels × 1024 pixels. For quantitative analysis of the albumin-bound Evans Blue, Z-stacks retinal images (step size 1.16 μm) of different regions of the vascular tree were acquired. To remove the eye artefacts caused by the sample procedures, projections of the middle Z-plane images were analysed with the magic wand tool (tolerance 120, eight pixels connected) of Fiji software (https://fiji.sc ).
Other immunohistochemistry analyses
GLP-1 and glutamate–aspartate transporter (GLAST) were evaluated by fluorescence microscopy using specific antibodies, as previously reported .
The quantification of glutamate was performed by reverse-phase ultraperformance liquid chromatography (Acquity UPLC; Waters, Milford, MA, USA) as aminoquinoline derivatives (AccQ-Tag Chemistry, MassTrak AAA method and instruments; Waters), following the methodology previously described by Narayan et al .
Data are presented as means ± SD. Comparisons of continuous variables were performed using ANOVA and Student’s t test with SPSS software (SPSS/Windows version 18; SPSS, Chicago, IL, USA). The statistical significance level was set at p < 0.05.
DPP-IV mRNA levels and protein concentration in the human retina
DPP-IV concentrations in human vitreous fluid
DPP-IV was also detected in the vitreous fluid but in a range 100-fold less than in plasma. No differences in either plasma or intravitreal DPP-IV levels were observed between type 2 diabetic and non-diabetic participants (plasma 497 ± 133 ng/ml vs 491 ± 151 ng/ml, respectively, p = NS; vitreous fluid 6.2 ± 8.32 ng/ml vs 5.9 ± 6.28 ng/ml, respectively, p = NS).
DPP4 mRNA levels and protein concentration in db/db mouse retina
As in the human samples, we did not find any significant differences in the mRNA and protein levels of DPP-IV between the neuroretinas from diabetic and non-diabetic mice (Fig. 1c, d). We were unable to obtain accurate data regarding the RPE because of difficulty in separating it from the choroid. In addition, the limited volume of vitreous fluid that can be obtained from db/db mice prevented the measurement of DPP-IV levels.
Topical administration of DPP-IVi increases the ocular content of GLP-1 and its downstream target EPAC-1
Neurodegeneration was prevented in diabetic mice treated with DPP-IVi
Topical administration of DPP-IVi prevents the increase of glutamate induced by diabetes by inhibiting GLAST downregulation
Furthermore, we observed that the GLAST, the main glutamate transporter produced by Müller cells, was significantly decreased in the retinas of diabetic mice treated with vehicle in comparison with non-diabetic mice (Fig. 6c, d). In diabetic mice treated with saxagliptin or sitagliptin, this downregulation of GLAST induced by diabetes was prevented (Fig. 6c, d).
Topical treatment with DPP-IVi prevents electroretinogram abnormalities induced by diabetes
Topical administration of DPP-IVi prevents albumin leakage induced by diabetes
The beneficial effects of the topical administration of DPP-IVi were unrelated to blood glucose levels
Evolution of blood glucose and body weight of the different experimental groups to evaluate the effect of saxagliptin
Blood glucose (mmol/l)
db/db – vehicle
30.4 ± 4.8
32.0 ± 3.0
45.7 ± 2.1
47.6 ± 2.3
db/db – saxagliptin
29.9 ± 4.6
31.0 ± 4.1
46.3 ± 1.5
49.3 ± 1.5
8.8 ± 1.3
8.8 ± 1.5
27.5 ± 1.3
28.4 ± 2.1
Evolution of blood glucose and body weight of the different experimental groups to evaluate the effect of sitagliptin
Blood glucose (mmol/l)
db/db – vehicle
28.8 ± 4.6
30.2 ± 3.0
46.3 ± 2.0
49.5 ± 1.4
db/db – sitagliptin
30.2 ± 3.5
31.6 ± 2.7
48.6 ± 1.4
51.1 ± 1.6
8.6 ± 1.0
9.2 ± 1.3
28.0 ± 1.6
28.8 ± 1.4
In the present study, we have shown that both mRNA levels and protein content of DPP-IV are much lower in the retina than in the liver or bowel. Notably, DPP-IV levels in human vitreous fluid were around 100-fold lower than in plasma. These results could explain our previous observation regarding the capacity of GLP-1 to reach the retina after its topical administration and its neuroprotective actions through GLP-1R. However, the higher DPP-IV concentrations detected in the RPE of diabetic in comparison with non-diabetic participants could decrease the availability of GLP-1 for reaching the neuroretina. In fact, the drugs that reach the retina via the transscleral route, as is the case for GLP-1, are first challenged by the choroids and the RPE . Therefore, a therapeutic strategy combining the topical administration of GLP-1 and a DPP-IVi could be a new and more effective approach to treating diabetic retinopathy.
From the pathophysiological point of view, it should be noted that GLP-1 is synthesised by the retina and its production is lower in diabetic individuals . The local production of GLP-1 could be extremely important because of its potential autocrine and paracrine neuroprotective actions through GLP-1R. Therefore, the endogenous deficit of GLP-1 that exists in the diabetic retina could be involved in the development of neurodegeneration. Apart from its endogenous retinal source, GLP-1 could reach the retina from the systemic circulation by crossing the BRB.
At present, the proportion of locally derived and blood-borne GLP-1 that exists in the retina is unknown. Nevertheless, it seems reasonable to postulate that the preservation or enhancement of the retinal content of GLP-1 by preventing its degradation could be a new strategy for treating the early stages of diabetic retinopathy. The design of experiments addressed to test this hypothesis was the second aim of the present study, and we found that both DPP-IVi (saxagliptin and sitagliptin) administered as eye drops led to a significant increase in the intraretinal content of GLP-1, thus preventing neurodegeneration and vascular leakage in db/db mice.
The mechanisms by which GLP-1R activation exerts its neuroprotective and vasculotropic action are complex and have recently been revised . In the present study, we found that saxagliptin and sitagliptin were able to increase both mRNA and protein levels of EPAC-1, a downstream cAMP signalling mediator. As cAMP is the main downstream messenger of GLP-1, it can be deduced that the EPAC-1 enhancement was due to GLP-1R activation. Interestingly, EPAC-1 plays an important role in the maintenance of the endothelial barrier and neuronal functions [20, 21, 22, 23], and therefore its upregulation after DPP-IV inhibition indicates this pathway as a new mechanism contributing to the protective vasculotropic actions of GLP-1R activation.
However, the beneficial effects of saxagliptin and sitagliptin should not be exclusively attributed to an increase in GLP-1R activation. In fact, DPP-IVis themselves could activate unrelated downstream GLP-1R pathways that might be involved in neuroprotection. In this regard, Dietrich et al  recently reported that the DPP-IVi linagliptin has a neuroprotective effect in Caenorhabditis elegans, a model of neurodegeneration induced by high glucose levels in which GLP-1R is not produced. Experiments performed using experimental knockout models for GLP-1R would help us to elucidate this important issue. Nevertheless, our results suggest that preservation of the retinal content of GLP-1 (locally produced and serum derived) via topical administration of DPP-IVi seems a good strategy for treating diabetic retinopathy.
We used a topical rather than oral administration of DPP-IVis in order to avoid the confounding factor represented by their capacity to lower blood glucose levels when they are orally administered. As we have shown in the present study, topical administration of either saxagliptin or sitagliptin was unable to change blood glucose levels, and consequently their effects were not mediated by an improvement of the diabetic milieu. With this strategy, we have been able to ascertain the direct effect of DPP-IVi independently of the improvement in blood glucose levels. As occurred with GLP-1R agonists , the topical administration of DPP-IVi also prevented the downregulation of GLAST induced by diabetes, thus reducing the extracellular concentration of glutamate, and consequently ameliorating excitotoxicity and retinal neuron death. In addition, the prevention of functional impairment measured by electroretinography was very similar to that previously reported with GLP-1 . Taken together, these findings also support the notion that the main action of DPP-IV is mediated by an increase of retinal GLP-1 levels.
Nevertheless, as mentioned above, we cannot rule out that other mechanisms unrelated to GLP-1/GLP-1R could be activated. In this regard, it should be noted that G-protein-coupled receptor (GPcR) coagulation factor II receptor-like 1 (F2RL1, previously known as Par2), which is abundant in retinal ganglion cells, can be activated by DPP-IV . After stimulation, F2RL1 promotes angiogenesis and inflammation, and therefore could be an important target when treating diabetic retinopathy. In addition, it has been reported that interleukin-1 receptor antagonist (IL-1RA), a competitive antagonist of IL-1β receptors, is a substrate of DPP-IV . Therefore, the inhibition of DPP-IV could mitigate the deleterious role of IL-1β in the pathogenesis of diabetic retinopathy by preventing the cleavage of IL-1RA.
The experimental evidence on the effects of the oral administration of DPP-IVi in diabetic retinopathy is limited and controversial [29, 30, 31, 32]. Gonçalves et al  showed that sitagliptin prevented changes in the endothelial distribution of tight junction proteins and decreased nitrosative stress, the inflammatory state and cell death by apoptosis in Zucker Diabetic Fatty (ZDF) rats (a rat model of type 2 diabetes). In addition, sitagliptin allowed for a recovery of the number of CD34+ cells present in the bloodstream to levels similar to their number in control animals, while the decreased ability of endothelial progenitor cells (EPCs) to adhere to the retinal vessels that was induced by diabetes was reversed. Even more interestingly, the same authors demonstrated that sitagliptin alleviated the increased BRB permeability in streptozotocin (STZ)-induced diabetic rats, which could not be attributed to a normalisation of glycaemia . However, Lee et al  recently reported that the intraperitoneal administration of sitagliptin induced vascular leakage in the retinas of mice with either retinopathy of prematurity or STZ-induced diabetes. The main mechanism involved was the activation of the SDF-1α/CXR4/Src/VE-cadherin signalling pathway. Consequently, further research seems warranted to clarify whether DPP-IVis exert beneficial or deleterious effects in the diabetic retina and also whether this is related to the different routes of administration.
To the best of our knowledge, there has been only one clinical trial specifically addressing the effect of either GLP-1R agonists or DPP-IV in the eye . In this double-blind, placebo-controlled, crossover clinical trial performed in 50 type 2 diabetic individuals without diabetic retinopathy, 6 weeks of sitagliptin treatment normalised retinal capillary flow and increased vasodilatory capacity during flicker-light exposure. However, no specific studies have so far evaluated the effects of GLP-1R or DPP-IVi on the endpoints of diabetic retinopathy.
Recent results of the Cardiovascular and Other Long-term Outcomes with Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN-6) study  showed that semaglutide (a long-acting GLP-1 analogue with an extended half-life of approximately 1 week) reduced the rate of cardiovascular morbidity and mortality but increased the progression of diabetic retinopathy. This merits comment because it contravenes the experimental results showing the beneficial effects of GLP-1R agonists in the development of diabetic retinopathy [9, 35, 36, 37, 38, 39, 40]. In this study, an unexpectedly higher rate of severe complications of diabetic retinopathy (i.e. vitreous haemorrhage, blindness, or conditions requiring treatment with an intravitreal agent or photocoagulation) in the semaglutide arm were detected in comparison with placebo.
However, several important drawbacks should be mentioned. First, the SUSTAIN-6 study was aimed at assessing cardiovascular but not diabetic retinopathy outcomes. In fact, retinal photographs were not taken systematically at baseline, and therefore it is impossible to know whether the two arms (placebo and semaglutide) were matched in terms of the degree of diabetic retinopathy. This is a crucial point because the degree of diabetic retinopathy is the most important factor in determining evolution during follow-up. Another potential factor accounting for the results reported on diabetic retinopathy from SUSTAIN-6 is the deleterious effect of a rapid lowering of HbA1c. This could have been a relevant factor regarding individuals in the most severe baseline retinopathy category . However, this critical information was not available due to the absence of diabetic retinopathy grading. A comprehensive review regarding the limiting factors of this study, which prevent valid results for diabetic retinopathy outcomes from being drawn, has recently been published .
In conclusion, the topical administration of the DPP-IVi saxagliptin and sitagliptin is effective in preventing neurodegeneration and vascular leakage in the diabetic retina. These effects can be attributed to an enhancement of GLP-1, but other mechanisms unrelated to the prevention of GLP-1 degradation cannot be ruled out. These findings could pave the way for clinical trials testing this new approach, alone or in combination with GLP-1, in the treatment of early stages of diabetic retinopathy.
All relevant data are included in the article and/or the ESM files.
This study was supported by grants from the Ministerio de Economía y Competitividad (SAF2016-77784 and PI16/00541) and the Fundació Marató TV3.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
CH, PB, CS, JS, MV, XG, OS, MG and RS made substantial contributions to the conception and design of the study, the acquisition of data, or analysis and interpretation of the data. All authors drafted the article or revised it critically for important intellectual content. All authors approved the final version of the manuscript to be published. RS is the guarantor of this work.
- 38.Fan Y, Liu K, Wang Q, Ruan Y, Zhang Y, Ye W (2014) Exendin-4 protects retinal cells from early diabetes in Goto-Kakizaki rats by increasing the Bcl-2/Bax and Bcl-xL/Bax ratios and reducing reactive gliosis. Mol Vis 10:1557–1568Google Scholar